Intelligent Agents (IA's) are special program packets that can move in a network from one computer to another. A general discussion of IA's may be found in the following web-based publications: [http://]www.computer.org/concurrency/pd1999/pdf/p3080.pdf; [http://]agents.umbc.edu (AgentWeb: Publications and presentations: Recommended Papers); [http://]www.botspot.com/news/000327ecommerce.html. “Mobile agents”, representing an alternative term for IA's, and as contemplated in the cellular telephone industry, are discussed in [http://]smatpush.cs.hut.fi/SoftwareAgents/Seminarpapers/Mobile—Agents/Mobile—A gents.htm.
Essentially, and as can be appreciated from the references cited above, IA's are entities that are capable of performing intelligent tasks inside computers. For example, IA's can visit servers connected to a network in order to find security breaches in these servers. IA's can also be given tasks to visit home pages and collect advertisements that are posted on such home pages.
It is expected that, in the future, IA's will be widely available and will be capable of performing tasks now requiring manual prompting or intervention. For instance, it is conceivable that one could send an IA to “visit” several shopping sites on the Internet to undertake the buying of goods, given a set of parameters for conducting the shopping (e.g., desired items, price range, etc.). Similarly, an IA could be directed to process bank accounts, buy stocks, pay credit cards, etc.
It will be appreciated, though, that since IA's are usually in the form of programs, they can be copied, perhaps by criminal elements, in order to perform illegal or unauthorized tasks. The potential also exists, of course, for criminal elements to produce their own IA's for intruding other computers (e.g., to illegally access someone's bank account). Accordingly, IA's could potentially be produced by hackers in the form of viruses or worms and could, as such, enter computers from infected disks and/or programs.
Consequently, it would appear that as the use of IA's increases, there will be increased importance placed on preventative measures against the illegal or unauthorized use of IA's. Its presently contemplated that IA's have some dedicated arrangement for self-identification (e.g., with ping numbers, names, digital signatures). (For instance, self-identification is contemplated in [http://]smaprtpush.cs.hut.fi/SoftwareAgents/Seminarpapers/Mobile—Agents/Mobile—Agents.htm.) It is believed, however, that such measures would not be foolproof, since the possibility exists that a hacker could copy an IA and adequately assess its functions (for instance, essentially any program that is moving in a network could be copied using devices that copy the program bit-by-bit, much as when a pirate copies programs on CD-ROMs).
A need has thus been recognized in connection with providing more effective measures that improve IA authentication in comparison with previous efforts.